Robust Video Watermarking Based On DWT And Synchronization Codes

Digital watermarking technology is permanently embedded in the digital media,and it is the information hiding technology which does not affect the normal use of the source carrier information.The complexity of video watermarking is relatively high,the anti attack ability of the watermark is higher than that of the image watermark,and the security,robustness and embedding capacity of video watermarking are also different.In order to improve the video watermarking security and robustness,the algorithm based on feature value transform combined with discrete wavelet transform was proposed;In order to improve the video watermarking embedding rate and the imperceptibility,the watermark embedding can be solved by linear predictive coding and 2D-DWT multi resolution.To enhance robustness of video watermarking and reduce the complexity of the algorithm,the eigenvalue decomposition video watermarking algorithm was proposed.According to the size of the embedded watermark information,the video frame image is segmented adaptively,and a part generates the synchronization code to find the embedding watermark;And another part of the process: In order to remove the correlation between the pixels of the watermark image,the first use of Amold scrambling to pre process the watermark image.Then the original video sequence is decomposed into two layers by DWT,and the low frequency coefficients are transformed into EVD transform,and the watermark is embedded into the generated diagonal matrix quantization.When extracting the watermark we use synchronization code to find the watermarked video frames in the band,the IDWT decomposition of the band,the decomposition of low-frequency EVD transform,and then the watermark sequence in the transform quantization matrix can be extracted.Experimental results show that the proposed method is not sensitive to the watermark,the security of the watermark and the robustness of the watermark are improved,and the complexity of the algorithm is reduced.In order to further improve the robustness and imperceptibility of video watermark,combined with the two layer DWT decomposition,a DWT domain robust video watermarking algorithm based on linear predictive coding technique is proposed.First,the I-frame video sequence from a frequency domain is divided into two parts,one of the part for synchronization code is embedded,so in the watermark extraction can fast locate the watermark,and the another part of linear predictive coding processing to obtain the prediction error.Then the prediction error for each I-frame block into 8 ?8,after the treatment of square 2D-DWT two layer wavelet decomposition.The watermark is embedded into the low frequency coefficients of the two layer DWT decomposition.Using the wavelet decomposition of video DWT domain can effectively enhance the anti attack ability of video watermarking.The above two algorithms,the security of the watermark,robustness and visual perception are improved.But there are some shortcomings of the watermarking algorithm,and put forward some ideas for future work.